Tesseract OCR 與 IronOCR：使用 C# 擷取 PDF 文本

using Tesseract;

// Step 1: Convert PDF page to PNG (requires a separate PDF rendering library)
// This example assumes the scanned PDF has already been converted to an image
string imagePath = "document-scan.png";

// Step 2: Initialize Tesseract with the language data path
using var engine = new TesseractEngine(@"./tessdata", "eng", EngineMode.Default);

// Step 3: Load the image and run OCR
using var img = Pix.LoadFromFile(imagePath);
using var page = engine.Process(img);

// Step 4: Extract recognized text
string extractedText = page.GetText();
Console.WriteLine($"Confidence: {page.GetMeanConfidence()}");
Console.WriteLine(extractedText);

// Optional: retrieve word-level bounding boxes
using var iter = page.GetIterator();
iter.Begin();
do
{
    if (iter.TryGetBoundingBox(PageIteratorLevel.Word, out var bounds))
    {
        string word = iter.GetText(PageIteratorLevel.Word);
        Console.WriteLine($"Word: {word} at {bounds}");
    }
} while (iter.Next(PageIteratorLevel.Word));

Imports Tesseract

' Step 1: Convert PDF page to PNG (requires a separate PDF rendering library)
' This example assumes the scanned PDF has already been converted to an image
Dim imagePath As String = "document-scan.png"

' Step 2: Initialize Tesseract with the language data path
Using engine As New TesseractEngine("./tessdata", "eng", EngineMode.Default)

    ' Step 3: Load the image and run OCR
    Using img As Pix = Pix.LoadFromFile(imagePath)
        Using page As Page = engine.Process(img)

            ' Step 4: Extract recognized text
            Dim extractedText As String = page.GetText()
            Console.WriteLine($"Confidence: {page.GetMeanConfidence()}")
            Console.WriteLine(extractedText)

            ' Optional: retrieve word-level bounding boxes
            Using iter As ResultIterator = page.GetIterator()
                iter.Begin()
                Do
                    Dim bounds As Rect
                    If iter.TryGetBoundingBox(PageIteratorLevel.Word, bounds) Then
                        Dim word As String = iter.GetText(PageIteratorLevel.Word)
                        Console.WriteLine($"Word: {word} at {bounds}")
                    End If
                Loop While iter.Next(PageIteratorLevel.Word)
            End Using

        End Using
    End Using

End Using

using IronOcr;

// Initialize the OCR engine (built on optimized Tesseract 5)
var ocr = new IronTesseract();
ocr.Configuration.PageSegmentationMode = TesseractPageSegmentationMode.Auto;
ocr.Configuration.ReadBarCodes = true; // Detect barcodes and QR codes alongside text

// Load PDF directly - no image conversion required
using var input = new OcrInput();
input.LoadPdf("scanned-document.pdf", Password: "optional-password");

// Apply preprocessing for low-quality scans
input.DeNoise();              // Remove background noise from scanned paper
input.Deskew();               // Correct rotation from camera angle
input.EnhanceResolution(300); // Ensure adequate DPI for accurate recognition

// Extract text from all pages
OcrResult result = ocr.Read(input);

Console.WriteLine($"Confidence: {result.Confidence}%");
Console.WriteLine($"Pages: {result.Pages.Count()}");
Console.WriteLine(result.Text);

// Export results as a searchable PDF
result.SaveAsSearchablePdf("searchable-output.pdf");

Imports IronOcr

' Initialize the OCR engine (built on optimized Tesseract 5)
Dim ocr As New IronTesseract()
ocr.Configuration.PageSegmentationMode = TesseractPageSegmentationMode.Auto
ocr.Configuration.ReadBarCodes = True ' Detect barcodes and QR codes alongside text

' Load PDF directly - no image conversion required
Using input As New OcrInput()
    input.LoadPdf("scanned-document.pdf", Password:="optional-password")

    ' Apply preprocessing for low-quality scans
    input.DeNoise()              ' Remove background noise from scanned paper
    input.Deskew()               ' Correct rotation from camera angle
    input.EnhanceResolution(300) ' Ensure adequate DPI for accurate recognition

    ' Extract text from all pages
    Dim result As OcrResult = ocr.Read(input)

    Console.WriteLine($"Confidence: {result.Confidence}%")
    Console.WriteLine($"Pages: {result.Pages.Count()}")
    Console.WriteLine(result.Text)

    ' Export results as a searchable PDF
    result.SaveAsSearchablePdf("searchable-output.pdf")
End Using

using IronOcr;
using System.Drawing;

var ocr = new IronTesseract();

// Restrict character recognition to digits and currency symbols for financial docs
ocr.Configuration = new TesseractConfiguration
{
    WhiteListCharacters = "0123456789.$,",
    PageSegmentationMode = TesseractPageSegmentationMode.SingleColumn
};

// Load only the first two pages from a financial report
using var input = new OcrInput();
input.LoadPdfPages("financial-report.pdf", new[] { 0, 1 });

// Target a specific crop region, such as an invoice total field
var cropRegion = new CropRectangle(x: 100, y: 500, width: 400, height: 200);
foreach (var page in input.Pages)
    page.AddCropRegion(cropRegion);

OcrResult result = ocr.Read(input);

foreach (var page in result.Pages)
{
    Console.WriteLine($"Page {page.PageNumber}:");
    foreach (var paragraph in page.Paragraphs)
        Console.WriteLine($"  ({paragraph.Confidence}%) {paragraph.Text}");
}

Imports IronOcr
Imports System.Drawing

Dim ocr As New IronTesseract()

' Restrict character recognition to digits and currency symbols for financial docs
ocr.Configuration = New TesseractConfiguration With {
    .WhiteListCharacters = "0123456789.$,",
    .PageSegmentationMode = TesseractPageSegmentationMode.SingleColumn
}

' Load only the first two pages from a financial report
Using input As New OcrInput()
    input.LoadPdfPages("financial-report.pdf", {0, 1})

    ' Target a specific crop region, such as an invoice total field
    Dim cropRegion As New CropRectangle(x:=100, y:=500, width:=400, height:=200)
    For Each page In input.Pages
        page.AddCropRegion(cropRegion)
    Next

    Dim result As OcrResult = ocr.Read(input)

    For Each page In result.Pages
        Console.WriteLine($"Page {page.PageNumber}:")
        For Each paragraph In page.Paragraphs
            Console.WriteLine($"  ({paragraph.Confidence}%) {paragraph.Text}")
        Next
    Next
End Using